Format: ORAL

Guisan A. (1), Chevalier M. (2), Adde A. (1), Zarzo-Arias A. (3), Goicolea T. (3), Broennimann O. (1), Petitpierre B. (4), Scherrer D. (5), Rey P.-L. (1), Mateo R.G. (3)

Department of Ecology and Evolution, University of Lausanne, Lausanne, Switzerland

Species distribution models (SDMs) have become a major tool in biogeography, especially in the context of global change assessments. A key application of SDMs is to derive projections in time, e.g. under climate change. SDMs are also based on several assumptions, one being that to make such temporal projection, the full realized environmental niche must be captured. Yet, the niche has a high risk to be truncated if the geographic – and associated environmental – range used to fit the model is smaller than the range occupied by the species, leading to potential mistakes when making temporal projections. A solution to avoid such niche truncation is to build SDMs in a hierarchical nested way, from global to local. Such approaches have already been applied to SDMs, to model invasive species, to downscale large scale models, to support species conservation, or to improve climate change projections. They were recently implemented in the new N-SDM pipeline in R to derive predictions for large numbers of species. However, the importance of using such hierarchically-nested approach has remained rather hidden and confidential so far in the SDM literature. The aim of this talk is to present a review and synthesis of the developments up to now, and to identify solutions and remaining challenges.